Benjamin Teke created MAPREDUCE-7441:
----------------------------------------
Summary: Race condition in closing FadvisedFileRegion
Key: MAPREDUCE-7441
URL: https://issues.apache.org/jira/browse/MAPREDUCE-7441
Project: Hadoop Map/Reduce
Issue Type: Bug
Components: yarn
Affects Versions: 3.4.0
Reporter: Benjamin Teke
This issue is similar to the one described in MAPREDUCE-7095, just for
FadvisedFileRegion.transferSuccessful. There are warning messages when multiple
threads are calling the transferSuccessful method:
{code:java}
2023-05-25 08:41:57,288 WARN org.apache.hadoop.mapred.FadvisedFileRegion:
Failed to manage OS cache for
/hadoop/data04/yarn/nm/usercache/hive/appcache/application_1684916804740_8245/output/attempt_1684916804740_8245_1_00_001154_0_10003/file.out
EBADF: Bad file descriptor
at org.apache.hadoop.io.nativeio.NativeIO$POSIX.posix_fadvise(Native Method)
at
org.apache.hadoop.io.nativeio.NativeIO$POSIX.posixFadviseIfPossible(NativeIO.java:271)
at
org.apache.hadoop.io.nativeio.NativeIO$POSIX$CacheManipulator.posixFadviseIfPossible(NativeIO.java:148)
at
org.apache.hadoop.mapred.FadvisedFileRegion.transferSuccessful(FadvisedFileRegion.java:163)
at
org.apache.hadoop.mapred.ShuffleChannelHandler.lambda$sendMapOutput$0(ShuffleChannelHandler.java:516)
at
io.netty.util.concurrent.DefaultPromise.notifyListener0(DefaultPromise.java:590)
at
io.netty.util.concurrent.DefaultPromise.notifyListenersNow(DefaultPromise.java:557)
{code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: mapreduce-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: mapreduce-issues-h...@hadoop.apache.org
